Filters








1,111 Hits in 7.2 sec

[Paper] Semantic Concept Detection based on Spatial Pyramid Matching and Semi-supervised Learning

Yoshihiko Kawai, Mahito Fujii
2013 ITE Transactions on Media Technology and Applications  
We also propose a training framework based on semi-supervised learning that uses a small number of labeled data points as a starting point and generates additional labeled training data efficiently, with  ...  Analyzing video for semantic content is very important for finding the desired video among a huge amount of accumulated video data.  ...  In contrast, when the initial data set was used directly without performing semi-supervised learning to train classifiers, lower average precision of 0.021 was achieved.  ... 
doi:10.3169/mta.1.190 fatcat:lsqfd3zydffpbkvws5ibyeshba

Modeling semantic relations between visual attributes and object categories via dirichlet forest prior

Xin Chen, Xiaohua Hu, Zhongna Zhou, Yuan An, Tingting He, E.K. Park
2012 Proceedings of the 21st ACM international conference on Information and knowledge management - CIKM '12  
To alleviate the workload of manual supervision and labeling in image categorization process, we introduce a semi-supervised training framework using soft-margin semi-supervised SVM classifier.  ...  In this paper, we deal with two research issues: the automation of visual attribute identification and semantic relation learning between visual attributes and object categories.  ...  We are the learning process from 50 labeled images bounding box per class (i.e. 2500 image bounding boxes in total) for training, The semi-supervised SVM use 50 labeled bounding boxes and an addition of  ... 
doi:10.1145/2396761.2398428 dblp:conf/cikm/ChenHZAHP12 fatcat:hnsf3g3uwnb6najihunrxffq3q

Large Scale Distributed Semi-Supervised Learning Using Streaming Approximation [article]

Sujith Ravi, Qiming Diao
2016 arXiv   pre-print
Traditional graph-based semi-supervised learning (SSL) approaches, even though widely applied, are not suited for massive data and large label scenarios since they scale linearly with the number of edges  ...  We also provide a distributed version of the algorithm that scales well to large data sizes.  ...  Acknowledgements We thank Partha Talukdar for useful pointers to MAD code and Kevin Murphy for providing us access to the Freebase-Relation dataset.  ... 
arXiv:1512.01752v2 fatcat:o5obilqb4vfxtb5yh6aeiaqdny

Teaching the Machine to Explain Itself using Domain Knowledge [article]

Vladimir Balayan, Pedro Saleiro, Catarina Belém, Ludwig Krippahl, Pedro Bizarro
2020 arXiv   pre-print
., LIME, and SHAP) produce explanations that are very hard to understand for non-data scientist persona.  ...  Machine Learning (ML) has been increasingly used to aid humans to make better and faster decisions.  ...  Acknowledgements The project CAMELOT (reference POCI-01-0247-FEDER-045915) leading to this work is cofinanced by the ERDF -European Regional Development Fund through the Operational Program for Competitiveness  ... 
arXiv:2012.01932v1 fatcat:igzh3ll45bbfzcc5smfgldywiu

Intelligent Evidence-Based Management for Data Collection and Decision-Making Using Algorithmic Randomness and Active Learning

Harry Wechsler, Shen-Shyang Ho
2011 Intelligent Information Management  
The constructive and all encompassing active learning (AL) methodology, which mediates and supports the above theme, is context-driven and takes advantage of statistical learning, in general, and semi-supervised  ...  Active learning employs explore and exploit actions characteristic of closed-loop control for evidence accumulation in order to revise its prediction models and to reduce uncertainty.  ...  This is characteristic of semi-supervised learning and is sometimes referred to as "hallucinations."  ... 
doi:10.4236/iim.2011.34018 fatcat:k36vb7cu4vcmbegzkgx7tk7ite

Consensus of Regression for Occlusion-Robust Facial Feature Localization [chapter]

Xiang Yu, Zhe Lin, Jonathan Brandt, Dimitris N. Metaxas
2014 Lecture Notes in Computer Science  
After localization, the occlusion state for each landmark point is estimated using a Gaussian MRF semi-supervised learning method.  ...  Recently, regression-based approaches to localization have produced accurate results in many cases, yet are still subject to significant error when portions of the face are occluded.  ...  A graph-based semi-supervised learning is also utilized to explicitly detect the occlusion.  ... 
doi:10.1007/978-3-319-10593-2_8 fatcat:osxmhrwmtjemvkmai2jxblt2t4

A Survey of Label-noise Representation Learning: Past, Present and Future [article]

Bo Han, Quanming Yao, Tongliang Liu, Gang Niu, Ivor W. Tsang, James T. Kwok, Masashi Sugiyama
2021 arXiv   pre-print
Therefore, it is urgent to design Label-Noise Representation Learning (LNRL) methods for robustly training deep models with noisy labels. To fully understand LNRL, we conduct a survey study.  ...  Classical machine learning implicitly assumes that labels of the training data are sampled from a clean distribution, which can be too restrictive for real-world scenarios.  ...  During the semi-supervised learning phase, they leverage variants of co-training, such as co-refinement on labeled data and co-guessing on unlabeled data. Specifically, Li et al.  ... 
arXiv:2011.04406v2 fatcat:76np6wyzvvag7ehy23cwyzdozm

RoGAT: a robust GNN combined revised GAT with adjusted graphs [article]

Xianchen Zhou, Yaoyun Zeng, Hongxia Wang
2021 arXiv   pre-print
Graph Neural Networks(GNNs) are useful deep learning models to deal with the non-Euclid data. However, recent works show that GNNs are vulnerable to adversarial attacks.  ...  Secondly, RoGAT further tunes the features to eliminate feature's noises since even for the clean graph, there exists some unreasonable data.  ...  Here we consider the semi-supervised node classification problem. Only parts of node V p = {v 1 , v 2 , · · · , v m } are part of labels with annotations.  ... 
arXiv:2009.13038v2 fatcat:gdbdw7aj4fgqdoy6hzlok2xoxm

Object Classification in Images of Neoclassical Furniture Using Deep Learning [chapter]

Bernhard Bermeitinger, André Freitas, Simon Donig, Siegfried Handschuh
2016 IFIP Advances in Information and Communication Technology  
It strives to deliver tools for analyzing the spread of aesthetic forms which are considered as a cultural transfer process.  ...  A data-driven bottom-up research routine in the Neoclassica research framework is the main use-case.  ...  By introducing Deep Neural Network Models from Machine Learning (ML) to this field, we hope that in particular semi-supervised learning methods will uncover clusters that were previously unknown.  ... 
doi:10.1007/978-3-319-46224-0_10 fatcat:mwix4yi3njh45hyaimaau3oeoq

Learning Scene-specific Object Detectors Based on a Generative-Discriminative Model with Minimal Supervision [article]

Dapeng Luo, Zhipeng Zeng, Nong Sang, Xiang Wu, Longsheng Wei, Quanzheng Mou, Jun Cheng, Chen Luo
2018 arXiv   pre-print
Here the human labeled training data or a generic detector are not needed.  ...  comparable performance to robust supervised methods, and outperforms the state of the art self-learning methods under varying imaging conditions.  ...  Universities Young Teacher Promotion Program-Outstanding Youth Foundation, China University of Geosciences (Wuhan)(CUGL170210), Fundamental Research Funds for National University, China University of  ... 
arXiv:1611.03968v4 fatcat:6o2xcfztifd2vmsct7jo26ae7a

Data-Driven Shape Analysis and Processing

Kai Xu, Vladimir G. Kim, Qixing Huang, Evangelos Kalogerakis
2016 Computer graphics forum (Print)  
Data-driven methods are also able to learn computational models that reason about properties and relationships of shapes without relying on hardcoded rules or explicitly programmed instructions.  ...  In contrast to traditional approaches that process shapes in isolation of each other, data-driven methods aggregate information from 3D model collections to improve the analysis, modelling and editing  ...  Acknowledgements We thank Zimo Li for proofreading this survey and the anonymous reviewers for helpful suggestions. Kalogerakis gratefully acknowledges support from NSF (CHS-1422441).  ... 
doi:10.1111/cgf.12790 fatcat:q76sq2syjvce5fjjjb7yoafww4

A survey on data‐efficient algorithms in big data era

Amina Adadi
2021 Journal of Big Data  
less training data and in particular less human supervision.  ...  AbstractThe leading approaches in Machine Learning are notoriously data-hungry.  ...  (iii) In multitask semi-supervised learning, tasks based their predictions on labeled data as well as unlabeled data.  ... 
doi:10.1186/s40537-021-00419-9 fatcat:v4uahsvhlzdldlxqf24bshmja4

Contrastive Code Representation Learning [article]

Paras Jain, Ajay Jain, Tianjun Zhang, Pieter Abbeel, Joseph E. Gonzalez, Ion Stoica
2021 arXiv   pre-print
For downstream semantic understanding tasks like summarizing code in English, these representations should ideally capture program functionality.  ...  ContraCode pre-trains a neural network to identify functionally similar variants of a program among many non-equivalent distractors.  ...  For these problems, learned code representations should be similar for functionally equivalent programs and dissimilar for non-equivalent programs (Figure 1).  ... 
arXiv:2007.04973v3 fatcat:bpqzjwtoebhh3in5q7qkk4sggq

Rekall: Specifying Video Events using Compositions of Spatiotemporal Labels [article]

Daniel Y. Fu, Will Crichton, James Hong, Xinwei Yao, Haotian Zhang, Anh Truong, Avanika Narayan, Maneesh Agrawala, Christopher Ré, Kayvon Fatahalian
2019 arXiv   pre-print
learned approaches) and to rapidly retrieve video clips for human-in-the-loop tasks such as video content curation and training data curation.  ...  To write these queries, we have developed Rekall, a library that exposes a data model and programming model for compositional video event specification.  ...  Acknowledgments We thank Jared Dunnmon, Sarah Hooper, Bill Mark, Avner May, and Paroma Varma for their valuable feedback. We gratefully acknowledge the support of DARPA under Nos.  ... 
arXiv:1910.02993v1 fatcat:grfxvproyrc5pmetcx62vhceui

Data-Driven Shape Analysis and Processing [article]

Kai Xu, Vladimir G. Kim, Qixing Huang, Evangelos Kalogerakis
2015 arXiv   pre-print
In addition, they are able to learn models that reason about properties and relationships of shapes without relying on hard-coded rules or explicitly programmed instructions.  ...  In contrast to traditional approaches, a key feature of data-driven approaches is that they aggregate information from a collection of shapes to improve the analysis and processing of individual shapes  ...  Active learning is a special case of semi-supervised learning in which a learning algorithm interactively queries the user to obtain desired outputs for more data points related to shapes.  ... 
arXiv:1502.06686v1 fatcat:upajios4y5a6dgf2zw7faqai4a
« Previous Showing results 1 — 15 out of 1,111 results